Published on Web 06/11/2003

Tremorgenic Indole Alkaloids. The Total Synthesis of (-)-Penitrem D Amos B. Smith, III*, Naoki Kanoh, Haruaki Ishiyama, Noriaki Minakawa, Jon D. Rainier, Richard A. Hartz, Young Shin Cho, Haifeng Cui, and William H. Moser Contribution from the Department of Chemistry, Monell Chemical Senses Center, and Laboratory for Research on the Structure of Matter, UniVersity of PennsylVania, Philadelphia, PennsylVania 19104

Received February 24, 2003; E-mail: [email protected]

Abstract: A convergent, stereocontrolled total synthesis of the architecturally complex tremorgenic indole alkaloid (-)-penitrem D (4) has been achieved. Highlights of the synthesis include an efficient, asymmetric synthesis of the western hemisphere; the stereocontrolled assembly of the I-ring; discovery of a novel autoxidation to introduce the C(22) tertiary hydroxyl group, required for tremorgenic activity; union of fully elaborated eastern and western hemispheres, exploiting an indole synthetic protocol developed expressly for this purpose; and a late-stage, stereoselective construction of the A and F rings exploiting a

Sc(OTf)3-promoted reaction cascade. The longest linear sequence leading to (-)-penitrem D (4) was 43 steps.

Ergot fungi produce a wide variety of architecturally complex natural products possessing significant bioregulatory properties. Among these, the indole-diterpenes pose significant economic problems for the livestock industry.1 Members of this family of environmental toxins are responsible for Dallisgrass poisoning (also called “paspalum staggers”),1a ryegrass staggers,1b-d and other related neurological disorders known to occur sporadically in the Southwest United States, Australia, and New Zealand. Domestic sheep and cattle that graze on ergot-infected grasses suffer from motion-induced tremors, limb weakness, ataxia, convulsions, and eventual death. Fortunately, the symptoms are reversible.1 From the perspective of human disorders, the symptoms presented by the livestock are similar to those associated with Wilson’s disease, multiple sclerosis, Parkinson’s disease, and epilepsy and are thereby suggestive of a possible Penitrem A was first isolated by Wilson and co-workers in 4a common neurochemical mechanism.2 1968 from the ergot fungus Penicillium cyclopium. Subse- quently, Steyne and co-workers reported the complete structures In addition to like biological profiles, the indole-diterpenes of the penitrems, isolated from the closely related Penicillium share common architectural features, the signature elements crustosum.3b,c,4b,c The absolute configuration was established by being an indole core, biosynthetically derived from tryptophan; the partial resolution method of Horeau.5 To date, no X-ray a diterpene framework arising from mevalonate; and the trans structures have been achieved, although we and others have vicinal quaternary methyl substituents.3 Among the indole attempted to obtain suitable crystals. Importantly, all six tremorgens, the penitrems [A-F(1-6)] represent the most complex members of the indole-diterpene alkaloid family. (3) (a) Acklin, W.; Weibel, F.; Arigoni, D. Chimia 1977, 31, 63. (b) de Jesus, A. E.; Hull, W. E.; Steyn, P. S.; van Heerden, F. R.; Vleggaar, R.; Wessels, P. L. J. Chem. Soc., Chem. Commun. 1982, 837-838. (c) de Jesus, A. E.; (1) (a) Cole, R. J.; Dorner, J. W.; Lansden, J. A.; Cox, R. H.; Pape, C.; Cunfer, Gorst-Allman, C. P.; Steyn, P. S.; van Heerden, F. R.; Vleggaar, R.; B.; Nicholson, S. S.; Bedell, D. M. J. Agric. Food Chem. 1977, 25, 1197- Wessels, P. L.; Hull, W. E. J. Chem. Soc., Perkin Trans. 1 1983, 1863- 1201. (b) Mantle, P. G.; Mortimer, P. H.; White, E. P. Res. Vet. Sci. 1978, 1868. 24,49-56. (c) di Menna, M. E.; Mantle, P. G. Res. Vet. Sci. 1978, 24, (4) (a) Wilson, B. J.; Wilson, C. H.; Hayes, A. W. Nature 1968, 220,77-78. 347-351. (d) Mantle, P. G.; Day, J. B.; Haigh, C. R.; Penny, R. H. Vet. (b) de Jesus, A. E.; Steyn, P. S.; van Heerden, F. R.; Vleggaar, R.; Wessels, Rec. 1978, 103, 403. (e) Springer, J. P.; Clardy, J. Tetrahedron Lett. 1980, P. L.; Hull, W. E. J. Chem. Soc., Chem. Commun. 1981, 289-291. (c) de 21, 231-234. (f) Gallagher, R. T.; Clardy, J.; Wilson, B. J. Tetrahedron Jesus, A. E.; Steyn, P. S.; van Heerden, F. R.; Vleggaar, R.; Wessels, P. Lett. 1980, 21, 239-242. (g) Dorner, J. W.; Cole, R. J.; Cox, R. H.; Cunfer, L.; Hull, W. E. J. Chem. Soc., Perkin Trans. 1 1983, 1847-1856. B. M. J. Agric. Food Chem. 1984, 32, 1069-1071. (5) (a) Horeau, A. Tetrahedron Lett. 1961, 506-512. (b) Horeau, A.; (2) Brimblecombe, R. W.; Pinder, R. M. Tremorgens and Tremorgenic Agents; Sutherland, J. K. J. Chem. Soc., Org. 1966, 247-248. (c) Herz, W.; Kagan, Scientechnia; Bristol, U.K., 1972; pp 14-15. H. B. J. Org. Chem. 1967, 32, 216-218.

8228 9 J. AM. CHEM. SOC. 2003, 125, 8228-8237 10.1021/ja034842k CCC: $25.00 © 2003 American Chemical Society Tremorgenic Indole Alkaloids ARTICLES penitrems bear a tertiary hydroxyl at C(22), the requisite Scheme 1 structural element required for the neurotoxicity.1g Penitrem D (4), the simplest member of the penitrem class, also possesses a cyclobutane ring and an eight-membered oxocane cyclic ether, but is devoid of the epoxide and chlorine substituents found in the more complex members of the penitrem family. The structural complexity of the indole diterpenes, in conjunction with their tremorgenic activity, first captured our attention in the early 1980s and has since led to the first total syntheses of (-)-paspaline (7),6a,b (+)-paspalicine (8),6c,d (+)- paspalinine (9),6c,d and quite recently (-)-21-isopentenylpaxilline (10).6e Armed with this experience, we embarked on the total synthesis of (-)-penitrem D (4). Several strategies were explored.6f-n Herein, we describe a full account of this program, which culminated in the first, and to date only, total synthesis of a member of the penitrem family.6n,7

Initial Synthetic Plan. From the outset we desired a convergent approach (Scheme 1). Retrosynthetically this entailed removal of the C(22) hydroxyl and refunctionalization of the C(11)-exo olefin to afford advanced intermediate 11, which we envisioned could serve as a common precursor for members of the penitrem family. For this plan to be successful, a viable the C(16) hydroxyl to complete ring A.8 We reasoned that this tactic would be required to introduce the C(22) hydroxyl. cascade of reactions might be possible in one operation. The Continuing this analysis, opening of the tetrahydropyran, the requisite eastern hemisphere 14, known as the Nolen-Sprengeler oxocane, and the F rings next led to a 2-substituted indole (12), lactone,6k,l was readily available in our laboratory, having been which could be further dissected to reveal three subtargets (13, designed as a common advanced intermediate for the synthesis 14, and 15). In the synthetic direction, the 2-substituted indole of several indole-diterpene tremorgens.9 To construct the I ring, (12) was anticipated to arise via union of a benzylic anion we anticipated attachment of the C(24) side chain via a Stork generated from 13 with an electrophile such as lactone 14, metalloenamine alkylation10 of the hydrazone derived from 14 followed by ring closure. Rings A and F would then be with epoxide 15. Stereocontrolled cyclization would then constructed via oxidation of the primary hydroxyl to an generate the tetrahydropyran.11 At the outset, the precise order aldehyde, execution of an intramolecular cyclization to furnish of indole construction and elaboration of the I ring was unclear. ring F and in turn a gramine-like fragmentation and capture by As will be presented, this flexibility of strategic events would (6) (a) Smith, A. B., III; Mewshaw, R. J. Am. Chem. Soc. 1985, 107, 1769- prove central to the eventual successful completion of the 1771. (b) Mewshaw, R. E.; Taylor, M. D.; Smith, A. B., III. J. Org. Chem. penitrem D synthesis. 1989, 54, 3449-3462. (c) Smith, A. B., III; Sunazuka, T.; Leenay, T. L.; Kingery-Wood, J. J. Am. Chem. Soc. 1990, 112, 8197-8198. (d) Smith, To implement this synthetic plan, we required (1) a viable A. B., III; Kingery-Wood, J.; Leenay, T. L.; Nolen, E. G.; Sunazuka, T. J. 2-substituted indole synthesis; (2) methods to construct the A, Am. Chem. Soc. 1992, 114, 1438-1449. (e) Smith, A. B., III; Cui, H. Org. Lett. 2003, 5, 587-590. (f) Smith, A. B., III; Visnick, M. Tetrahedron F, and I rings; and (3) a protocol to insert the C(22) hydroxyl. Lett. 1985, 26, 3757-3760. (g) Smith, A. B., III; Visnick, M.; Haseltine, - Although preliminary accounts describing these methods have J. N.; Sprengeler, P. A. Tetrahedron 1986, 42, 2957 2969. (h) Haseltine, - J.; Visnick, M.; Smith, A. B., III. J. Org. Chem. 1988, 53, 6160-6162. (i) appeared,6f n their critical importance to the overall development Smith, A. B., III; Haseltine, J. N.; Visnick, M. Tetrahedron 1989, 45, 2431- 2449. (j) Smith, A. B., III; Ohta, M.; Clark, W. M.; Leahy, J. W. of the penitrem D program demands their brief introduction here. Tetrahedron Lett. 1993, 34, 3033-3036. (k) Smith, A. B., III; Nolen, E. G.; Shirai, R.; Blase, F. R.; Ohta, M.; Chida, N.; Hartz, R. A.; Fitch, D. (8) For examples of gramine-type fragmentations, see: Brewster, J. H.; Eliel, M.; Clark, W. M.; Sprengeler, P. A. J. Org. Chem. 1995, 60, 7837-7848. E. L. Org. React. 1953, VII,99-197. (l) Smith, A. B., III; Hartz, R. A.; Spoors, P. G.; Rainier, J. D. Isr. J. Chem. (9) Lactone (+)-14 can be prepared from commercially available (-)-Wieland- 1997, 37,69-80. (m) Smith, A. B., III; Kanoh, N.; Minakawa, N.; Rainier, Miescher ketone in 16 steps and 8% overall yield. J. D.; Blase, F. R.; Hartz, R. A. Org. Lett. 1999, 1, 1263-1266. (n) Smith, (10) (a) Stork, G.; Benaim, J. J. Am. Chem. Soc. 1971, 93, 5938-5939. (b) A. B., III; Kanoh, N.; Ishiyama, H.; Hartz, R. A. J. Am. Chem. Soc. 2000, Stork, G.; Benaim, J. Org. Synth. 1977, 57,69-72. 122, 11254-11255. (11) A similar protocol had also proven effective in our paspalicine and (7) Rivkin, A.; Nagashima, T.; Curran, D. P. Org. Lett. 2003, 5, 419-422. paspalinine synthetic ventures.

J. AM. CHEM. SOC. 9 VOL. 125, NO. 27, 2003 8229 ARTICLES Smith et al.

New Indole Synthesis, Model Studies to Construct the A, converted to the o-nitroselenide,14 treated with camphorsulfonic F, and I Rings, and a Protocol to Introduce the C(22) acid in benzene, and subjected to oxidative-elimination of the Hydroxyl. Since the earliest annals of natural products chem- seleno group.15 Pleasingly, the hexacyclic ABCDEF analogue istry, bioactive indole alkaloids have evoked enormous interest of penitrem D (e.g., 22), possessing spectroscopic data that in the development of new synthetic approaches to the indole correlated well with that of (-)-penitrem D (4), was obtained ring.12 In 1985, based on the above penitrem D synthetic in 38% for the three steps. Structural confirmation of 22 was scenario, we designed an expedient method for the regiospecific achieved by single-crystal X-ray analysis of the bromo derivative synthesis of mono- and disubstituted indoles (Scheme 2).6f,g This 23. protocol involves treatment of N-trimethylsilyl-o-toluidine or a To model a possible tactic for elaboration of the I ring of related derivative with 2.2 equiv of n-BuLi to generate the penitrem D, we prepared acetate 27 (Scheme 4), exploiting the lithium dianion 17, which upon acylation with either esters or Stork metalloenamine alkylation of the hydrazone10 derived from lactones leads to an intermediate (18) capable of intramolecular (+)-24 with epoxide (-)-25.6j Camphorsulfonic acid-promoted heteroatom Peterson olefination13 to furnish the desired substi- cyclization, followed by removal of the MOM group with 1 M tuted indole 19; yields are in general good (50-98%). HCl in methanol, provided cis-tetrahydropyran (+)-28, along with what was thought to be the minor trans isomer (ca. 3.4: Scheme 2 1). The structure of (+)-28 was again secured by X-ray crystallographic analysis.

Scheme 4

To test the feasibility of constructing the A and F rings in a single operation via a cyclization-gramine-like fragmentation- cyclization cascade, we selected the ABCDEF hexacycle 22 in racemic form as an early target.6h,i Indole 20 was therefore prepared from aniline 13,6h oxidized to the corresponding aldehyde, and subjected to camphorsulfonic acid in methanol (Scheme 3). The resultant mixed methyl aminal 21 was next

Scheme 3

The fact that formation of the A and F rings in 22 and the I ring in (+)-28 proceeded with good stereoselectivity fortuitously under the identical mild acidic conditions (i.e., camphorsulfonic acid/benzene) proved highly seductive, suggesting the intriguing possibility of a combined “one-flask” A, F, and I ring con- struction.16 As will become apparent, this prospect markedly influenced the evolution of the penitrem D synthetic program. The remaining synthetic issue prior to committing to the penitrem D venture concerned installation of the C(22) hydroxyl. A third model study was performed.6j Oxidation of (+)-28 with singlet oxygen employing methylene blue as sensitizer in

(12) For a review on the synthesis of indole derivatives, see: Sundberg, R. J. Indoles; Academic Press Inc.: San Diego, 1996. (13) Kru¨ger, C.; Rochow, E. G.; Wannagat, U. Chem. Ber. 1963, 96, 2132- 2137. (14) Grieco, P. A.; Gilman, S.; Nishizawa, M. J. Org. Chem. 1976, 41, 1485- 1486. (15) Sharpless, K. B.; Young, M. W. J. Org. Chem. 1975, 40, 947-949. (16) This cascade of reactions is affectionately referred to as the “Shazam” approach, a term to the best of our knowledge first employed by Eaton to describe unsuccessful attempts by several research groups to effect the Lewis acid-promoted isomerization of C20H20 and C22H24 polycycloalkanes to dodecahedrane and dimethyl dodecahedrane, see: Eaton, P. E. Tetra- hedron 1979, 35, 2189-2223, and references therein.

8230 J. AM. CHEM. SOC. 9 VOL. 125, NO. 27, 2003 Tremorgenic Indole Alkaloids ARTICLES

20 CHCl3, followed by NaBH4 reduction of the resulting hydro- Semmler-Wolff aniline aromatization. The sequence proved peroxide, furnished (-)-29 (54% yield, two steps), as the sole highly efficient, requiring 13 steps and furnishing (-)-13 in 17% isolable product (Scheme 5). The yield for this transformation overall yield. could be improved to 63% using the sensitizer hematoporphyrin Union of the Eastern and Western Hemispheres and IX, in pyridine, followed by reduction with PPh3 prior to Elaboration of an ABCDEFGH Octacyclic Model. With both isolation.17 The stereogenicity of the newly generated hydroxyl aniline (-)-13 and the protected Nolen-Sprengeler lactone (-)- was again secured by single-crystal X-ray analysis. 336k,l in hand, we explored their union via the aforementioned indole synthesis (Scheme 7).6m Treatment of the N-silylated Scheme 5 derivative of (-)-13 with 2 equiv of s-BuLi, followed by addition of lactone (-)-33 and work up involving exposure of the intermediate product to silica gel, provided indole (-)-34 in near quantitative yield.21 Parikh-Doering oxidation22 of the primary hydroxyl, and in turn treatment with methanolic HCl to remove simultaneously the hydroxyl protecting groups and ketal, and then selective reinstallation of the TIPS silyl group furnished (+)-35 as a single isomer (stereochemistry undefined) in 41% yield for three steps.

Scheme 7

Initiation of the Penitrem Synthetic Venture: Construc- tion of Western Aniline (-)-13. Having achieved experimental support for the strategic transformations required by our synthetic plan, we initiated work on (-)-penitrem D (4) with the synthesis of 13 possessing the requisite absolute stereo- chemistry (Scheme 6), employing a sequence essentially identi- cal to that of our earlier racemic synthesis of 13, which was used to access the ABCDEF model 22.18 Highlights of this sequence include an effective four-step photochemical genera- tion of (-)-31 from (-)-30,6k which proceeded in 49% overall yield, a Woodward-Wilds19 modification of the Robinson annulation to append the carbon skeleton for ring D, and a

Scheme 6

At this juncture, we decided to evaluate the tandem AF ring construction tactic with the more complex substrate (+)-35.An alternative would have been to introduce the I ring first (vide infra). As in the earlier model study, treatment with camphor- sulfonic acid in benzene furnished the octacyclic enone (-)-36 in 39% yield based on recovered (+)-35 (Scheme 8). The structure of (-)-36 was confirmed by X-ray crystallography. “One-Pot” Formation of the A, F, and I Rings. Having achieved access to the advanced ABCDEFGH model system [(-)-36], we turned to the possibility of the “one-pot” construc-

(20) (a) Semmler, W. Chem. Ber. 1892, 25, 3352-3354. (b) Wolff, L. Annalen 1902, 322, 351-391. (c) For a review of the transformation, see: Conley, R. T.; Chosh, S. In Mechanism of Molecular Migrations; Thyagarajin, B. (17) See the Supporting Information for the experimental procedure. S., Ed.; Interscience: New York, 1971; Vol. 4, pp 251-308. (18) Hartz, R. A. Ph.D. Thesis, University of Pennsylvania, 1996. (21) With the fully elaborated coupling partners, the heteroatom Peterson (19) (a) Shunk, H.; Wilds, A. L. J. Am. Chem. Soc. 1949, 71, 3946-3950. (b) olefination did not go to completion without external promoters such as Woodward, R. B.; Sondheimer, F.; Taub, D.; Heusler, K.; McLamore, W. silica gel or heating; see also ref 6e. M. J. Am. Chem. Soc. 1952, 74, 4223-4251. (22) Parikh, J. R.; Doering, W. E. J. Am. Chem. Soc. 1967, 89, 5505-5507.

J. AM. CHEM. SOC. 9 VOL. 125, NO. 27, 2003 8231 ARTICLES Smith et al.

Scheme 8 alcohol, and removal of the MTM group, then furnished (+)- 39, embodying the complete carbon skeleton and functionality required to attempt the proposed A, F, and I ring construction in a single operation.25 Initial observations were encouraging. Treatment of (+)-39 with camphorsulfonic acid in benzene did indeed result in the required reaction cascade, furnishing the A and F rings, and in the formation of the tetrahydropyran I ring (Scheme 10). The stereochemical outcome was, of course, the question. After extensive NMR analysis, we were forced to conclude that, although formation of rings A and F had proceeded correctly, as predicted by model (-)-36, the stereogenicity at C(28) in the I ring was incorrect vis-a`-vis the penitrem skeleton. This result was clearly unexpected given the earlier successful I ring model study (see Scheme 4).6m

Scheme 10 tion of the A, F, and I rings. Stork metalloenamine alkylation10 of the dimethylhydrazone derived from (+)-35 with epoxide (-)-15 as the electrophile (Scheme 9), followed by protection of the newly generated hydroxyl (PivCl, DMAP), furnished (+)- 38, albeit in 27% yield for the three steps.23 In our hands, a particularly troubling (and recurring) aspect of the Stork protocol is the necessity to carry out the reaction on a reasonable scale

Scheme 9

The I Ring Model Study Revisited. After considerable analysis, we recognized that the major product obtained by initial treatment of model 27 (ca. 3.4:1 mixture) with camphorsulfonic acid was not the desired cis-tetrahydropyran (Scheme 11), but instead the trans isomer (+)-41, the result of a kinetically

Scheme 11

when advanced (+)-35 was in limited supply. To circumvent this problem, we added the readily available auxiliary hydrazone 37.6j We surmised that the auxiliary hydrazone would also facilitate anion equilibration, required for successful Stork metalloenamine alkylation. The use of 37 did improve the yield + 24 (23) Attempts to use a MOM protecting group at C(25) as in the earlier model of ( )-38 to 38%. Hydrolysis of the hydrazone, followed in study failed due to our inability to differentiate the secondary and tertiary turn by reduction of the enone, acetylation of the resultant hydroxyls at C(25) and C(16), respectively.

8232 J. AM. CHEM. SOC. 9 VOL. 125, NO. 27, 2003 Tremorgenic Indole Alkaloids ARTICLES controlled cyclization, and that (+)-28, the thermodynamically not be feasible and that independent approaches to the A, F, more stable isomer, for which we had structural confirmation and I ring systems would have to be achieved. The successful by X-ray analysis, had arisen during the 1 N HCl treatment to construction of the three rings in (-)-40 nonetheless was taken remove the MOM protecting group. This scenario was easily as a milestone in our penitrem D venture. confirmed both by conversion of (+)-43 to the corresponding Penitrem D: A Revised Synthetic Plan. To preserve the MOM ether (+)-41 and by isomerization of (+)-43 to (+)-28 convergent strategy, we chose to move the AF ring construction monitored by NMR.26 Unfortunately, all attempts to effect a tactic that had proven so successful in the early model similar isomerization of the advanced penitrem D ABCDEF- studies6h,i,m to late in the synthetic sequence. This strategic GHI nonacycle [(-)-40] to access the correct penitrem core reorganization suggested 45 (Scheme 12), possessing a prein- (11), via HCl treatment, proved unsuccessful due to the inherent acid lability of the A ring oxocane. Scheme 12 Given the critical nature of the required isomerization, we thought it prudent to explore other solvent and acid systems. Again, initial studies proved encouraging. Treatment of 44 with a catalytic amount of several acids including the Grieco solvent system27 (Table 1), followed by removal of the benzoyl moiety,

Table 1. Tetrahydropyran Formation

entry conditions yield for two steps 1 0.25 equiv CSA, benzene, 4 h 52% (1:1) 2 3 M LiClO4-Et2O, 5 °C f rt, 23 h no reaction 3 0.01 equiv AcOH, 3 M LiClO4-Et2O, 85% (1:1) 5 °C f rt, 18 h 4 0.01 equiv CSA, 3 M LiClO4-Et2O, 88% (8:1) 5 °C f rt, 16 h 5 0.01 equiv HClO4, 3 M LiClO4-Et2O, 84% (13:1) ° f 0 C rt, 1.5 h stalled tetrahydropyran I ring, to be a more viable advanced intermediate. A similar 2-substituted indole synthesis employing furnished (+)-28. The highest level of selectivity for the desired western hemisphere aniline 46, with a fully elaborated eastern cis-tetrahydropyran (+)-28 (ca. 13:1) was obtained with 0.01 hemisphere 47, would then be required. A Second-Generation Eastern Hemisphere (47). To gener- equiv of HClO4. These conditions however failed to generate the requisite A and F rings with (+)-35. Taken together these ate 47, we first explored installation of the C(23) side chain - - 6k,l results forced the conclusion that construction of the A, F, and onto the now familiar Nolen Sprengeler lactone ( )-14, 10 I rings, employing a single set of reaction conditions, would again via a Stork metalloenamine alkylation. The lactone moiety however precluded direct Stork alkylation with epoxide (24) Without the auxiliary hydrazone 37, we were unable to detect the desired (-)-15. Lactone (-)-14 was therefore converted to a mixture coupling product between (+)-35 and (-)-15. + + (25) Intermediates (+)-38 and (+)-39 were both single isomers; the stereo- of acetals ( )-48a and ( )-48b via a three-step sequence chemistry was not determined. (Scheme 13). The acetals, after separation, were individually (26) Heating (e.g., 80 °C) was required for isomerization in benzene due to the lower polarity of the solvent compared to that of methanol. Due to the converted to the corresponding dimethyl hydrazones. Stork overlap of signals from residual MeOH, water, and compounds (+)-43 and - + metalloenamine alkylation with epoxide ( )-15, employing ( )-28, we decided to utilize benzene-d6 instead of CD3OD. 6j (27) Grieco and co-workers exploited these conditions in an elegant synthesis again the auxiliary hydrazone (+)-37, followed by benzoylation of yuehchukene, see: (a) Grieco, P. A.; Clark, J. D.; Jagoe, C. T. J. Am. of the derived hydroxyls led to (+)-49a and -b. In this case the Chem. Soc. 1991, 113, 5488-5489. (b) Henry, K. J., Jr.; Grieco, P. A. J. Chem. Soc., Chem. Commun. 1993, 510-512. yields were somewhat improved to 58% and 76% respectively, compared to (+)-38 (see Scheme 9). Hydrolysis of both the hydrazones and methyl acetals in turn, followed by PDC oxidation, furnished lactone (+)-50. For the purposes of material advancement, the overall sequence could, of course, be carried forward with the mixture of 48a and 48b. To elaborate the I ring, we explored the conditions that had proven successful with model system 44 (vide supra). Lactone 51 was prepared from (+)-50 by 1,2-reduction of the enone,

J. AM. CHEM. SOC. 9 VOL. 125, NO. 27, 2003 8233 ARTICLES Smith et al.

Scheme 13 observed was of considerable significance. To implement this tactic, we first removed the MTM protecting group in (+)-50 (Table 2; entry 1) and then subjected the resultant alcohol to the conditions of TMSOTf and Et3SiH; surprisingly diene (-)- 53, the product of elimination, as opposed to silane reduction was the exclusive product. Replacement of TMSOTf with TfOH led to the desired cis-tetrahydropyran (+)-52 in 38% yield for the two steps, albeit still accompanied by the diene in 45% yield (entry 2). Importantly, none of the trans-isomer was observ- ed. A one-pot procedure comprising MTM removal, cyclization, and silane reduction proved feasible with Et3SiH-TfOH, to furnish a mixture (1:1.8) of (+)-52 and (-)-53, respectively (entry 3). Although we could transform (-)-53 to (+)-52 in 35% yield [82% yield based on recovered (-)-53],30 employing the same Et3SiH-TfOH conditions, we sought further optimiza- tion. Eventually, use of toluene as solvent in place of CH2Cl2, with careful control of reaction temperature, proved effective, furnishing the cis-tetrahydropyran (+)-52 in 60% yield.

Table 2. Reductive Cyclization Leading to (+)-52

acetylation of the resultant hydroxyl, and removal of the MTM protecting group (Scheme 14). Treatment of 51 with HClO4 yield (%) (0.01 equiv) in 3 M LiClO4-ether furnished a mixture of cis- entry conditions (+)-52 (−)-53 and trans-tetrahydropyrans 52. Unfortunately, the yield (32%) - - and stereoselectivity (cis:trans ) 1:1) of the ring closure proved 1 (1) MeI, CaCO3,CH3CN THF H2O (4:1:1), 082 55 °C incommensurate with material advancement.28 Clearly, an (2) 1.2 equiv TMSOTf, 9.7 equiv Et3SiH, alternate approach to the I ring was required. CH2Cl2,0°C 2 (1) MeI, CaCO3,CH3CN-THF-H2O (4:1:1), 38 45 Scheme 14 55 °C (2) 4.8 equiv TfOH, Et3SiH-CH2Cl2 (1:1), 0 °C 3 4.9 equiv TfOH, Et3SiH-CH2Cl2 (1:1), 35 63 -78 °C f -20 °C 4 8.4 equiv TfOH, Et3SiH-toluene (1:1), 60 0 -40 °C f rt

To complete construction of the advanced lactone (+)-54, the benzoyl group was removed (methanolic K2CO3), followed by regeneration of the partially hydrolyzed lactone with 1-(3- dimethylaminopropyl)-3-ethylcarbodiimide hydrochloride (ED- CI);31 the overall yield for the two steps was 90% (Scheme 15). In 1989 Nicolaou reported an elegant synthesis of oxepanes Scheme 15 via reductive cyclization of simple hydroxy ketones through the intermediacy of an oxonium ion.29 The conditions of choice were Et3SiH and TMSOTf. That complete syn diastereoselectivity was

(28) The major byproduct was a diene i (32% yield).

Introduction of the C(22) Hydroxyl. Having arrived at the

(29) Nicolaou, K. C.; Hwang, C.-K.; Nugiel, D. A. J. Am. Chem. Soc. 1989, complete carbon skeleton of the eastern hemisphere, we turned 111, 4136-4137. to introduce the C(22) hydroxyl employing the photo-oxidation

8234 J. AM. CHEM. SOC. 9 VOL. 125, NO. 27, 2003 Tremorgenic Indole Alkaloids ARTICLES conditions developed with model (+)-28 (see Scheme 5). Scheme 17 Unfortunately, all attempts to achieve this transformation with the advanced system (+)-54 proved totally unrewarding (Table 3). We speculate, as illustrated, that the vicinal quaternary

Table 3. Photo-Oxidation of (+)-54

entry conditions 1 1 O2, hν, py, 35 °C, hematoporphyrin IX; PPh3 1 2 O2, hν, benzene, reflux, hematoporphyrin IX, TPP; PPh3 1 3 O2, hν, CHCl3,35°C, methylene blue; NaBH4, MeOH 1 4 O2, hν, MeOH-benzene (1:1), 35 °C, methylene blue; NaBH4, MeOH 1 5 O2, hν, MeOH-benzene (1:1), 35 °C, methylene blue; PPh3 (TPP ) 5,10,15,20-tetraphenyl-21H,23H-porphyrin) methyl groups block the appropriate trajectory of singlet oxygen, possible with (+)-28, to both faces of the tetrasubstituted olefin.

At this juncture, we recalled that during our earlier syntheses of (+)-paspalicine and (+)-paspalinine, a model â,γ-unsaturated ketone (+)-56 had proven highly susceptible to autoxidation, is not the rate-limiting step. Instead, we suggest that silica gel- quantitatively affording the corresponding hydroperoxide (+)- promoted keto-enol tautomerization, to generate the C(22- 57 (Scheme 16).6d Similar autoxidation of steroids has been 24) dienol, may be the rate-limiting step.34 The alpha stereo- selectivity of the autoxidation process presumably derives from Scheme 16 stereoelectronic factors, namely, a “product-like” transition state wherein direct chair formation (pathway a) is favored as opposed to a twist boat (pathway b) as illustrated in Scheme 17.35 To arrive at the fully elaborated eastern hemisphere, all that remained was stereoselective reduction of the C(25) carbonyl and protection of the secondary hydroxyl as the TES ether. Several reducing agents were screened; L-selectride produced attributed to hydrogen atom abstraction, followed by the addition the highest level of stereoselectivity (Table 4). Protection of of oxygen to the resultant radical.32 To explore this possibility, the hydroxyl completed the synthesis of the eastern hemi- (+)-54 was oxidized to the corresponding â,γ-unsaturated sphere(-)-47, the structure being confirmed by single-crystal ketone (+)-60 with the Dess-Martin periodinane.33 During the X-ray analysis. purification of (+)-60, we observed a minor amount of hydroperoxide formation. This observation led us to develop a three-step oxidation/autoxidation/reduction protocol employing in turn the Dess-Martin periodinane, air/silica gel oxidation, and PPh3 reduction to furnish (-)-59 in 60% yield for the three steps (Scheme 17). In an effort to accelerate the oxidation, the radical initiator AIBN was added. No improvement in the reaction rate occurred, suggesting that hydrogen atom abstraction Fragment Union. Recognizing the potential acid sensitivity of the intermediates anticipated to arise post 2-substituted indole (30) Higher reaction temperature and/or longer reaction time resulted in over- construction, we chose to replace the MOM protecting group reduction of the olefin in the side chain. (31) Dhaon, M. K.; Olsen, R. K.; Ramasamy, K. J. Org. Chem. 1982, 47, 1962- in (-)-13 with the more easily removable TMS group. Not 1965. (32) Nickon, A.; Mendelson, W. L. J. Org. Chem. 1965, 30, 2087-2089. surprisingly, the harsh conditions required to remove the MOM (33) Dess, D. B.; Martin, J. C. J. Am. Chem. Soc. 1991, 113, 7277-7287. group also led to loss of the TBS group. Reprotection was

J. AM. CHEM. SOC. 9 VOL. 125, NO. 27, 2003 8235 ARTICLES Smith et al.

Table 4. Reduction of (-)-59 Scheme 19

yield (%) entry conditions (47:63)

1 NaBH4, CeCl3, MeOH, 0 °C 96 (1:8) 2 Li(O t-Bu)3AlH, THF, 0 °C f rt 94 (1:2) 3 NaBH(OAc)3, AcOH, 80 °C f 100 °C no reaction End Game. The last major hurdle to overcome was genera- 4 L-Selectride, THF, -78 °C 86 (19:1) tion of the A and F rings (Scheme 20). Parikh-Doering oxidation22 of the primary hydroxyl followed by selective achieved with a TIPS group at the primary hydroxyl36 and a removal of the TMS and TES groups afforded a serviceable TMS at the C(16) hydroxyl to furnish (+)-46 [48% yield, 100% equilibrium mixture of 64, 65a, and 65b (3:1:3). Initial attempts yield based on recovered starting material (Scheme 18)]. to construct the A and F rings with CSA in benzene provided the desired oxocane (-)-66, albeit in low yield (ca. 19%), along Scheme 18 with a considerable amount of the elimination product (-)-67 (56%). The structure of (-)-67 was deduced via 2D NMR experiments. In an attempt to improve the ratio of (-)-66 and (-)-67, a series of Lewis acids was screened; best results were 37 obtained with 1.2 equiv of Sc(OTf)3 in benzene. These conditions furnished (-)-66 in 62% yield. The requisite â-ster-

Scheme 20

Following our 2-substituted indole protocol, o-toluidine (+)- 46 was converted to the N-trimethylsilyl-o-toluidine derivative and, in turn, treated with 2.1 equiv of s-BuLi. Addition of lactone (-)-47 furnished the coupled product, which upon exposure to silica gel underwent the requisite heteroatom Peterson olefina- tion to provide indole (-)-45 (Scheme 19).6n The excellent yield (e.g., 81% yield) was quite gratifying.

(34) A similar autoxidation of an enol (ii) was observed by Kuwajima and co- workers during the synthesis of (+)-taxusin, see: Hara, R.; Furukawa, T.; Kashima, H.; Kusama, H.; Horiguchi, Y.; Kuwajima, I. J. Am. Chem. Soc. 1999, 121, 3072-3082.

(35) (a) House, H. O.; Tefertiller, B. A.; Olmstead, H. D. J. Org. Chem. 1968, 33, 935-942. (b) Paquette, L. A.; Belmont, D. T.; Hsu, Y.-L. J. Org. Chem. 1985, 50, 4667-4672. (c) Stereoselective Alkylation Reactions of Chiral Metal Enolates. Evans, D. A. Asymmetric Synthesis; Morrison, J. D., Ed.; Academic Press: New York, 1983; Vol. 3, Chapter 1, pp 1-110. (36) The lability of the TBS protecting group suggested that the use of the more robust TIPS ether would prove advantageous.

8236 J. AM. CHEM. SOC. 9 VOL. 125, NO. 27, 2003 Tremorgenic Indole Alkaloids ARTICLES eochemistry at C(18) of (-)-66, as in the earlier model studies, manner. The initial approach afforded nonacycle (-)-40, was again obtained with excellent selectivity (>95:5 by 1H possessing the undesired stereogenicity at C(28) via a dramatic NMR). one-pot construction of the A, F, and I rings. The second- With the critical A and F rings secure, final conversion to generation strategy called for the construction of the fully penitrem D (4) proved straightforward. Acetylation of the C(25) elaborated eastern hemisphere (-)-47, possessing a preinstalled hydroxyl (Scheme 21), followed by removal of the TIPS group, I ring; highlights of the I ring construction included a Stork furnished alcohol (-)-68, which in turn was subjected to Grieco metalloenamine alkylation, an acid-promoted reductive cycliza- selenation14 of the primary hydroxyl followed by oxidative tion, and autoxidation at C(22). Union of the fully elaborated elimination to install the C(11) exo olefin.15 Saponification then eastern and western hemispheres was then achieved employing provided (-)-penitrem D (4), identical in all respects to a sample the 2-substituted indole synthesis, developed expressly for the of the natural product (e.g., 1H and 13C NMR, IR, HRMS, and (-)-penitrem D synthetic venture. The excellent efficiency of chiroptic properties). this union (81%) stands as a testimony for the application of this method in complex molecule synthesis. Finally, a Sc(OTf)3- Scheme 21 promoted cationic cascade completed construction of the A-I nonacyclic system. The longest linear sequence to (-)-penitrem D(4) required 43 steps from (-)-Wieland-Miescher ketone. Acknowledgment. Support was provided by the National Institutes of Health (Institute of General Medical Sciences) through grant GM-29028. Postdoctoral fellowships are gratefully acknowledged by N.K. (the Japan Society for the Promotion of Science for Young Scientists), H.I. (Uehara Memorial Founda- tion), and J.D.R. and W.H.M. (The National Institutes of Health). We are also indebted to Dr. Mike Kaufmann and Professor Keiji Tanino (Hokkaido University) for helpful discussions on the I ring formation and autoxidation, respectively, and Christopher M. Adams for help with the preparation of this manuscript. In addition, we thank Drs. George Furst, Rakesh Kohli, and Patrick Carroll, Directors of the University of Pennsylvania Spectro- scopic Facilities, for assistance in obtaining high-field NMR spectra, high resolution mass spectra, and X-ray crystallographic analysis, respectively. Supporting Information Available: Experimental procedures and analytical data for compounds (-)-13,(+)-28-(-)-36,(+)- 38-(+)-41,(+)-43,(-)-45-(+)-50,(+)-52-(+)-54,(-)-59, (+)-60,(+)-62,(+)-63,(-)-66-(-)-68. This material is available free of charge via the Internet at http://pubs.acs.org. JA034842K - (37) (a) Kobayashi, S.; Hachiya, I.; Araki, M.; Ishitani, H. Tetrahedron Lett. Summary. The first total synthesis of ( )-penitrem D has 1993, 34, 3755-3758. (b) Kobayashi, S.; Hachiya, I.; Ishitani, H.; Araki, been achieved in a highly convergent and stereocontrolled M. Synlett 1993, 472-474.

J. AM. CHEM. SOC. 9 VOL. 125, NO. 27, 2003 8237 本文献由“学霸图书馆-文献云下载”收集自网络，仅供学习交流使用。

学霸图书馆（www.xuebalib.com）是一个“整合众多图书馆数据库资源，

提供一站式文献检索和下载服务”的24 小时在线不限IP 图书馆。 图书馆致力于便利、促进学习与科研，提供最强文献下载服务。

图书馆导航：

图书馆首页 文献云下载 图书馆入口 外文数据库大全 疑难文献辅助工具